Application of symbolic inductive learning methods to gene expression analyses

This paper deals with application of selected symbolic inductive learning methods, as well as feature selection and classifier combining methods, to some real gene expressions data. We show that for this class of data, it is possible to improve system performance remarkably, by simultaneous application of different methods of gathering information from attribute space, especially through feature selection and combination of various classifiers. All results are obtained from knowledge mining system WEKA and our original system EMPIRIC.